日本統計学会誌

ASYMMETRIC ASYMPTOTIC CONFIDENCE REGION FOR A MULTIVARIATE LINEAR STRUCTURAL RELATIONSHIP

Yoshiko Isogawa

1982 年 12 巻 2 号 p. 97-104
発行日: 1982年
公開日: 2009/01/22

DOIhttps://doi.org/10.11329/jjss1970.12.97

ジャーナルフリー

抄録を表示する抄録を非表示にする

For the problem of finding a confidence region for a multivariate linear structural relation, Okamoto & Isogawa [7] proposed a method to construct an asymptotic confidence region as the sample size tends to infinity by treating the true variables symmetrically. This article presents a method to construct another asymptotic confidence region by treating the variables asymmetrically.

抄録全体を表示

PDF形式でダウンロード (274K)
UNIFORMLY MINIMUM VARIANCE UNBIASED ESTIMATION OF QUANTILES AND PROBABILITIES FOR THE PARETO DISTRIBUTION OF THE FIRST KIND

Noriaki Setô, Kosei Iwase

1982 年 12 巻 2 号 p. 105-112
発行日: 1982年
公開日: 2009/01/22

DOIhttps://doi.org/10.11329/jjss1970.12.105

ジャーナルフリー

抄録を表示する抄録を非表示にする

The uniformly minimum variance unbiased estimators of the quantile and the probability in the Pareto distribution of the first kind are constructed. The variance of the former is obtained in an analytic form, and for some cases its values are calculated numerically.

抄録全体を表示

PDF形式でダウンロード (252K)
THE DISTRIBUTION OF THE RANDOM MANHATTAN DISTANCE IN A HYPERCUBE AND ITS APPLICATIONS

Osamu Sugano

1982 年 12 巻 2 号 p. 113-124
発行日: 1982年
公開日: 2009/01/22

DOIhttps://doi.org/10.11329/jjss1970.12.113

ジャーナルフリー

抄録を表示する抄録を非表示にする

The distribution of the k-dimensional random Manhattan distance in a k-dimensional unit hypercube is given and some test statistics based on it are proposed for testing k-dimensional randomness of a given sequence of numbers. Properties of the proposed statistics are investigated and it is shown by numerical experiments that our test statistics are uncorrelated with the traditional ones.

抄録全体を表示

PDF形式でダウンロード (438K)
A DECISION FUNCTION FOR SELECTING SUBPOPULATIONS ON THE BASIS OF A TRUNCATED SAMPLE

Kazuo Noda

1982 年 12 巻 2 号 p. 125-132
発行日: 1982年
公開日: 2009/01/22

DOIhttps://doi.org/10.11329/jjss1970.12.125

ジャーナルフリー

抄録を表示する抄録を非表示にする

In a selection problem, the optimum selection φ_??_ depends on an unknown parameter θ. We propose φ_??_ as a substitute for φ_??_, where _??_ is a BAN estimator for θ based on a subsidiary sample of size n. The asymptotic risk of φ_??_ as n tends to infinity is evaluated under several regularity conditions.

抄録全体を表示

PDF形式でダウンロード (320K)
D-OPTIMAL DESIGNS OF RESOLUTION V IN SOME CLASS OF FRACTIONAL 2^m FACTORIAL DESIGNS

Masahide Kuwada

1982 年 12 巻 2 号 p. 133-140
発行日: 1982年
公開日: 2009/01/22

DOIhttps://doi.org/10.11329/jjss1970.12.133

ジャーナルフリー

抄録を表示する抄録を非表示にする

This paper presents the determinant of the covariance matrix of the estimates of effects based on a fractional 2^m factorial (2^m-FF) design T of resolution V, where T is constructed by adding some restricted assemblies to an orthogonal array (O-array) or by removing some restricted ones from an O-array of index unity. In the class of 2^m-FF designs considered here, we also present Doptimal designs of resolution V for each m with 4_??_m_??_6 and for a range of practical values of N, where N denotes the total number of assemblies. In addition, these results are compared with the corresponding D-optimal designs of resolution V in the class of balanced fractional 2^m factorial designs.

抄録全体を表示

PDF形式でダウンロード (329K)
A DEFINITION OF NBU PROBABILITY MEASURES

Fumio Ohi, Toshio Nishida

1982 年 12 巻 2 号 p. 141-151
発行日: 1982年
公開日: 2009/01/22

DOIhttps://doi.org/10.11329/jjss1970.12.141

ジャーナルフリー

抄録を表示する抄録を非表示にする

This paper presents a definition of NBU probability measures on ((R⁺)^m, (B⁺)^m), where R⁺=[0, ∞) and B⁺ is the class of Borel subsets of R⁺, and it is shown that our NBU probability measures have several desirable properties, for example, they are closed under the formation of coherent systems, they are closed under the limits in distribution and so on. Boundary probability measures of our class of NBU probability measures are given explicitly. Finally we present some examples of NBU probability measures, especially the bivariate Erlang distributions are shown to be NBU.

抄録全体を表示

PDF形式でダウンロード (392K)
絶対偏差和を最小にする多項式のあてはめ

河原靖

1982 年 12 巻 2 号 p. 153-170
発行日: 1982年
公開日: 2009/01/22

DOIhttps://doi.org/10.11329/jjss1970.12.153

ジャーナルフリー

抄録を表示する抄録を非表示にする

2次元変量に関するデータ(x_i, y_i) (i=1, 2, …, n)に対して,絶対偏差和が最小になるような多項式曲線の満足すべき条件を明らかにし,この条件を満たす曲線を探索することにより,多項式曲線のあてはめができることを示す.また2次, 3次のような低次の多項式曲線のあてはめに対して,グラフを用いる技法を提示する.

抄録全体を表示

PDF形式でダウンロード (670K)
マクロエコノメトリック・モデルの特定化と外挿特性

浜田文雅

1982 年 12 巻 2 号 p. 171-178
発行日: 1982年
公開日: 2009/01/22

DOIhttps://doi.org/10.11329/jjss1970.12.171

ジャーナルフリー

抄録を表示する抄録を非表示にする

This paper attempts to make a comparison of the characteristics of time configurations of the dynamic solutions in the period of extrapolation of a Japanese quarterly macroeconometric model KOM for the case where the investment equations are replaced with drastically simplified forms in their specification. It was found that the specification change has brought about a remarkable change in the time configurations of solutions in the extrapolations.

抄録全体を表示

PDF形式でダウンロード (438K)
日本統計学会第50回大会記録

1982 年 12 巻 2 号 p. 179-193
発行日: 1982年
公開日: 2009/01/22

DOIhttps://doi.org/10.11329/jjss1970.12.179

ジャーナルフリー

PDF形式でダウンロード (1039K)

J-STAGEへの登録はこちら（無料）